CV

Education, experience, awards, and service. The PDF download button above links to the latest CV.

Contact Information

Name Hoyong Kim
Professional Title Postdoctoral Researcher
Email hoyong.kim.96@gmail.com

Professional Summary

Postdoctoral researcher at the Visual & Geometric Intelligence Lab, Seoul National University. I study how deep networks organize information in their representation space (Neural Collapse, representation collapse) and extend this geometric view to the interpretability of vision-language-action models and world models for robotics.

Experience

  • 2026 - present

    Seoul, Republic of Korea

    Postdoctoral Researcher
    Visual & Geometric Intelligence Lab, Seoul National University
    Working with Prof. Jaesik Park on mechanistic interpretability of vision-language-action models and world models for robotics.
  • 2020 - 2021

    Gwangju, Republic of Korea

    Research Assistant
    Intelligence Representation & Reasoning Lab (IRR Lab), AI Graduate School, GIST
    • Analyzed the search space of neural networks and how restricting models to an effective search space helps training.
  • 2019 - 2020

    Daejeon, Republic of Korea

    Research Student
    Research Data Sharing Center, Division of National S&T Data, KISTI
    • Developed a clustering module for issue analysis from news articles, improving both speed and accuracy.
    • Built a gold-standard Korean synonym dataset from the Korean Open Dictionary and a Siamese network for synonym similarity.
  • 2015 - 2018

    Gunsan, Republic of Korea

    Research Intern
    Data Mining & Artificial Intelligence Lab (DILAB), Kunsan National University
    • Deep learning model for traffic accident point prediction from about 10,000 sensing records in Daegu, Korea.
    • Web document classification for people with the same name, e-mail record analysis, and sentiment analysis on news articles.

Education

  • 2021 - 2026

    Gwangju, Republic of Korea

    Ph.D.
    Gwangju Institute of Science and Technology (GIST)
    Artificial Intelligence
    • Advisor: Prof. Kangil Kim, Intelligence Representation & Reasoning Lab (IRR Lab)
    • Dissertation: From Theory to Remedy: Addressing Representation Collapse for Enhanced Class-Discriminative Clarity
  • 2018 - 2020

    Daejeon, Republic of Korea

    M.S.
    University of Science and Technology (UST), KISTI Campus
    Big-Data Science
    • Advisor: Prof. Dong-min Seo
    • Thesis: Study on Text Mining and Clustering for Issue Analysis on News Articles
  • 2014 - 2018

    Suwon, Republic of Korea

    B.S.
    Ajou University
    Software

Teaching

Awards

  • 2024
    RA Student Research Achievement Scholarship
    GIST AI Graduate School

    Awarded in 2023 and 2024.

  • 2022
    Excellence Prize, Graduate Student Research Poster Exhibition
    GIST AI Graduate School
  • 2020
    Minister's Commendation, 3rd Government Innovation Proposal Hackathon
    Ministry of the Interior and Safety, Republic of Korea
  • 2020
    Excellence Prize, Campus Patent Universiade
    Korea Invention Promotion Association (KIPA)
  • 2019
    Best Paper Award, ICCC 2019
    Korea Contents Association
  • 2018
    Best Paper Award, KCA 2018 Spring Conference
    Korea Contents Association
  • 2016
    Grand Prize, Ajou Creative Software Festival
    Ajou University

Academic Service

Reviewer: ICLR 2025, NeurIPS 2025, ICLR 2026, ICML 2026 (Gold Reviewer)

Skills

Programming: Python, PyTorch, C, Java, TensorFlow
Tools: Linux, Git, LaTeX, Vim
Background: Linear Algebra, Numerical Optimization, Machine Learning, Deep Learning, Computer Vision, Natural Language Processing, Data Mining

Interests

Research Interests: Neural Collapse, Representation Learning Theory, Mechanistic Interpretability of VLA Models, World Models, Imbalanced Learning, Continual Learning

Languages

Korean : Native speaker
English : Fluent